►
From YouTube: Kubernetes SIG Storage 20170608
Description
Kubernetes Storage Special-Interest-Group (SIG) Meeting - 06 June 2017
Meeting Notes/Agenda: https://docs.google.com/document/d/1-8KEG8AjAgKznS9NFm3qWqkGyCHmvU6HVl0sk5hwoAE/edit#heading=h.gr59dea4wwlw
Find out more about the Storage SIG here: https://github.com/kubernetes/community/tree/master/sig-storage
Moderator: Brad Childs (Red Hat)
Chat Log:
00:16:22 Luis Pabón: Is there a link to an issue so we know what the issues are with stabilization support in EBS
00:19:20 hekumar: so we already fixed plenty of issues with EBS recently. There is no one issue though
00:20:50 hekumar: https://github.com/kubernetes/kubernetes/issues/44656 the test I was talking about btw
00:21:11 hekumar: and https://github.com/kubernetes/kubernetes/issues/43300
A
Public
assess
if
the
snorts
thing
tonight,
I
gonna
walk
through
the
agenda.
Google
is
out,
they
are
on
an
off-site
and
inside
asked
me
to
run
this
place.
There's
a
couple
issues
on
our
agenda
and
we
also
need
to
go
through
our
laying
spreadsheet
for
one
seven
and
not
anything
off
it
made
it.
So
let
me
start
there
I
know.
Some
of
these
features
are
owned
by
people,
not
on
the
call.
So
every
little
sniffle
from
any
my
the
NFS
dredging
issued
yes
yeah.
No,
we.
B
C
A
B
E
F
A
D
F
A
B
A
A
H
A
H
I
I
A
I
Cloud
provider
Azure
there
was
there
may
have
been
some
refactor
that
was
tried
to
put
into
the
actual
disk
underlying
systems,
but
as
far
as
I
know,
we're
gonna
try
and
avoid
that.
So
we
basically
we're
gonna,
restart
work
on
this
and
just
been
rethink.
Keller
approaching
here
I
mean
woman
was
super
helpful
and
in
kind
of
shaping
the
thought
on
that.
So,
if
you
are
you
here,
woman,
if
so,
thank
you
very
much
for
yeah.
J
We
open
this
poor
request
week
or
two
ago,
and
we
originally
proposed
four
one
seven,
but
it's
gonna
take
a
little
more
work,
so
we're
we're
waiting
for
one
point
eight
to
get
this
in,
and
there
is
some
design
question
in
in
the
PR
something
session
Mossad
and
Liggett
and
I'm,
not
sure
yeah.
So
I
don't
know,
that's
probably
the
right
place
to
to
have
the
design
discussion,
but
I
just
want
to
get
it
some
attention.
J
As
far
as
the
you
know,
the
way
to
have
an
out
of
tree
plugin
now
communicate
with
it
within
tree.
Basically,
we
have
an
out
of
tree
controller
or
operator.
That's
watching
it.
So
we
want
to
have
that
communication
done
in
a
kubernetes,
natal
native
way,
whether
it's
teepee,
ours
and
or
TBR's
or
CRTs,
or
the
proposed
way.
J
A
J
D
J
A
K
This
sorry
guys
I
added
this
today,
I'm
sorry
I
was
talking.
I
wasn't
me.
This
is
Louise
I
added
this,
because
I
just
want
to
make
sure
that
that
the
sake
was
following
this
I
I
see
that
the
implementation
was
part
of
six
doors.
But
sorry
if
a
fossil
is
part
of
six
storage,
a
label,
but
the
implementation
I,
don't
see
it
being
followed
by
six
doors.
I
want
to
see
if
the
sake
was
going
to
follow
this
proposal,
because.
K
E
D
Thank
you
now,
as
I
guess
from
the
back.
So
that's
our
if
your
provision,
yes
and
we
want
to
attach
to
the
hosts
that
is
outside
of
the
AWS,
they
got
into
trouble,
but
the
first
in
the
first.
If
you
want
to
do
that,
the
card
provider
initialization
graph
failure,
because
if
I
wanted
provider
in
secure
cloud
provider
that
attacks
is
there
a
lot
of
yoga
AWS
on
us?
Otherwise
just
won't
even
starts
so
I.
Don't
think
that's
an
issue
at
all
this
kind
of
our
self
carried
the
cloud
provider
against
his
own.
L
Team,
so
basically
the
idea
was
from
from
our
experience.
It
opens
that
so
notice
that
there's
a
module
called
post
way
to
basically
serve
as
a
public
library
to
do
the
attach
and
detach
for
various
protocols
except
I,
ska,
CFC
and
etc.
So
we
are
thinking
about.
Maybe
bringing
these
feature
to
true
penalties
may
be.
The
two
blood
know
that
most
of
the
the
attach
detach
operations,
who
you
know
just
import,
this
library
and
and
then
basically,
everyone
who
use
it
because
also
functionalities
are
common
and
public
operations.
L
L
D
That's
a
good
idea,
sanitation
and
we're
over
statute.
So
last
week
we
got
found
hypervisors
and
one
way
for
hypervisor
than
one
with
a
cinder.
The
fabric
connection
is
the
hypervisor
library
really
hard
on.
You
know
we
don't
really
have
a
hypervisor
assume
us
to
do
it,
so
the
common
libraries
being
OS
free,
which
makes
sense
to
hyper
by
this,
may
not
make
sense
to
their
convenience
or
progress.
D
L
D
L
A
F
I'm
gonna,
take
it
from
my
Google
Doc
and
put
it
in
a
PR
today
all
case
event
for
this
document,
but
I
hope
everyone
can
attend.
I've
had
a
lot
of
feedback
from
the
community
around
it,
so
I
wanted
to
make
sure
we
were
on
track
and
on
scope,
I
sent
it
to
the
signaling.
Listen
to
the
invite!
Please
let
me
know
if
he
didn't
get
it.
A
F
A
A
For
the
early
stages
of
design,
I
think,
unless
you're
clearly
confident
that
this
design
is
exactly
how
it's
going
to
go.
I
if
you
are
then
I
can
make,
are
down
just
fine.
If
you
give
you
that
there's
some
wiggle
room
or
some
points
that
need
further
clarification
and
I
might
work
to
see
the
name
Google
Doc
for
now
either
way.
I.
F
Mean
yeah
gave
in
the
churn
of
every
single
design
decision
we've
made
on
the
sig
I.
Don't
think
anyone
can
say
with
any
confidence
that
they
think
that's
the
right
thing
out
of
the
gate.
I
think
we
have
a
lot
of
great
opinions
in
this
case,
so
I'll
just
leave
it
in
the
Google
Doc
then,
and
pass
that
out
after
this
meeting.
A
All
right,
so
our
boarding
dachshund
process.
This
is
another
thing
I'd,
like
the
you
know,
I'd,
like
probably
present,
for
when
we
discuss
just
so
that
we
don't
make
any
decisions
they're
adamantly
against,
for
whatever
reason
is
there
anything
why
to
discuss
and
kind
of
a
shallower
context
regarding
this
right
now
or
is
everyone?
Okay?
Putting
this
off
I
guess:
they're,
very
emergency
commissars.
You.
A
No
I'm,
not
against
it.
What
I'm
saying
is
like
it's.
The
onboarding
Docs
in
the
process
were
officially
I'm
boarding.
All
right,
I
would
I
want
to
make
sure
that
Google
is
part
of
the
discussion
of
that
process.
I
simply
pretty
much
own.
The
repo
so
I'm
a
little
hesitant
to
go
over
and
link
them
up
with
darts
and
process
and
then
show
it
to
them,
then
give
it
back
but
I'm
asking.
K
K
M
But
we
have
done
some
contribution
in
terms
of
any
to
end
tests
that
are
now
being
ported
to
AWC.
Understand
correctly.
So
so
what
we
are
selling
with
is
you
know
we
want
to
contribute
to.
You
know,
apart
from
just
being
willing
to
just
each
here
for
this
or
Mississippi
codebase.
We
want
to
continue
to
forty
when
it
is
so
I
would
like
to
know
more.
Maybe
I
can
take
it
offline.
You
know
how
we
can
you
know
what
is
on
board
of
what
is
on
body.
A
Right
so
the
amount
of
mean
side
I've
talked
about
this
with
Brian
grant
was.
If
there's
you
know
some
mass
amount
of
code
that
people
have
to
you
know
show
that
they've
contributing
in
that
they
know.
What's
going
on
and
I
say,
we've
reached
so
many
lines
and
cover
that
they
actually
got
committed.
Then
they
can
be
added
to
the
organization
young,
that
the
process
isn't
really
well-defined.
A
I
I
know
that
tests
and
bugs-
and
you
know
it's
probably
the
first
order-
I-
think
that
people
can
pick
up
the
you
know
actually
being
the
way
to
throw
that's
required
to
to
get
on
board.
But
we
haven't
come
up
with
those
docks
yet,
and
my
suspicion
is
that
the
this
bullet
point
is
here
for
us
to
come
up
with
some
kind
of
working
group
in
Morse.
We
define
that
okay.
M
A
A
M
D
K
The
only
thing
that
I
suggest
in
this
onboarding
process
is
that
we
consider
you
know
the
same
way
that
OpenStack
does
it
is
that
that
they're
called
core
reviewers
in
the
reviewing
is
sometimes
more
important
than
sending
the
code.
We
get
a
lot
of
code,
but
we
get
very
little
reviewers
right
and
so
I
would
say
that
if,
if
your
team
wants
to
participate
more
than
they
become,
you
know
reviewers
of
more
things.
F
Thank
you,
so
I
kind
of
misspoke
when
I
had
originally
added
this
to
the
document,
so
I
just
wanted
to
clarify
the
CN
CF
has
come
forward
with
the
desire
to
kind
of
have
testing
environments
so
that
we're
able
to
test
everything,
that's
entry
to
the
best
of
our
ability,
and
we
all
know
that
it's
been
a
constant
challenge
to
have
a
test
environment,
to
debug
things
and
to
test
things.
And
so
what
we're
trying
to
do
is
is
work
with
the
CN
CF
and
various
vendors
on.
F
You
know,
understanding
what
needs
to
be
done,
what
kind
of
hardware
that
were
required
and
how
we
would
share
that.
So
we
had
a
pretty
lengthy
discussion
around
getting
started
on
that
process.
With
that
we
came
to
the
consensus
and
I
sent
this
out
in
a
separate
storage,
sig
email.
You
know
anyone
contributing
code
to
kubernetes
core
and
wanting
to
have
their
there
plug
as
part
of
this
kind
of
federated
testing
environment.
You
know
we're
kind
of
looking
to
solicit
the
community
on
representatives
of
this,
not
just
code
but
really
time.
F
The
only
thing
that
will
make
it
successful
is
that
people
are
willing
to
put
time
forward
to
make
it
happen.
So,
if
you're
on
the
call
and
that's
something
you
desire-
and
you
represent
a
storage
vendor,
then
you
know
please
seek
me
out,
and
we
can
talk
about
it
more
in
length.
But
again,
this
is
just
that
the
infancy
we're
trying
to
really
figure
out
how
it's
going
to
work
and
how
it's
going
to
maintain
itself
long
term.
F
So
I'm
just
reporting
that,
as
part
of
the
status
moving
forward,
we
had
some
kind
of
stale
pr's
on
the
list
that
have
since
been
resolved
and
I've
actually
had
a
few
people
reach
out
to
me
since
last
week
that
are
also
interested
in
getting
on
board
to
help
out
with
testing
I.
Personally
think
this
is
a
a
great
way
to
contribute.
As
a
new
contributor
to
the
community
into
storage
is
through
tests,
you
learn
a
lot.
F
You
learn
a
lot
even
more
than
storage,
but
if
you're
doing
end-to-end
tests,
the
interaction
with
the
other
components
and
kubernetes
so
I
encourage
anyone
new
to
the
project.
To
please
reach
out
to
me,
and
you
know
we
can
try
to
find
something
appropriate
for
you
to
work
on
so
otherwise.
I
think
that's
where
we
are
with
testing.
F
There
is
a
standing,
maybe
I
don't
have
it
on
the
eye.
I
have
a
part
of
the
kubernetes
sig
Google
Groups.
There
is
a
meeting
every
other
week.
It's
called
a
tea
storage
test
meeting.
That
would
be
the
place
to
show
up
it's
not
included
in
this
meeting
because
we
generally
don't
have
enough
time.
But
if,
for
some
that
isn't
on
your
calendar,
please
let
me
know:
I
can
resend
the
invite
if
you
have
joined
that
community
after
the
fact
and
then
I
will
make
sure
I
have
it
on
the
cube
community
calendar
as
well.